At 08:10 PM 16-10-00 +0100, Richard Robinson wrote:
>On Mon, 16 Oct 2000, Bob Archer wrote:
>
>> To finish off with, I am going to restate my basic premise:
>>
>> The more variants of abc programs accept, the less useful abc is as an
>> exchange mechanism.
>
>Reductio ad absurdum: if abc programs accept no variants of abc they'll be
>universal exchange mechanisms.
>
>I don't have to re-edit (primarily) abc2win-generated abc because abc2ps
>_accepts_ the abc2win variant. I have to re-edit them because it
>_doesn't_.
>
>I think you must mean 'generate' ?
No, I meant accept, although the argument can also apply to generate. I
might not have explained it very well though - I was also thinking more of
human writers of abc than programs. I stated my full argument some time ago
on here and got no response, so I decided not to repeat it in that post.
I'll have another go.
I am going to assume that many people who use abc also use some piece of
software that reads abc. Whilst I am sure there are people out there who
just use abc to scribble down a tune on a bit of paper and then transfer it
directly to staff notation I am not considering them in this argument.
If a program allows a feature of abc to be used, users will use that
feature, whether or not that feature appears in the standard. Most users
probably won't care if that feature appears in the standard - all they want
is to get something working with their particular abc program. Some of
these users will put their abc up on the net, other users will download
that abc which might or might not work with their abc software. This leads
to exactly what we have now. A body of abc tunes, some of which use
features that are not in the standard and are not available on all abc
software.
My experience is that computer users will do whatever they are allowed to
get away with - I don't even mean this in a negative way, it is just the
way the world is. If the programs accept non-standard abc that's what we
will end up with.
My contention is essentially your Reductio ad absurdum, except I'm
expressing it as a sliding scale rather than as an absolute. As I said in
my previous post, there are two extremes - one where all software
implements the standard, no more and no less and the other where all of the
software ignores the standard completely. The first is impossible to
achieve, and is undesirable because of the lack of innovation allowed. The
second is undesirable because it makes abc unusable as an exchange mechanism.
We have to be at some point on the sliding scale between the two extremes,
and that involves balancing things off against each other. In particular we
seem to have "usefulness as an exchange mechanism" pitted directly against
"allows programmers to innovate" at the moment - the classic dilemma for
all standards.
To come back to your example, I would state it as "You have to edit
abc2win-generated abc because abc2win does not generate standard conforming
abc".
Again, I think that speed of updating the standard is critical to finding
the right balance point.
Bob
To subscribe/unsubscribe, point your browser to: http://www.tullochgorm.com/lists.html
